The Impact of Backpack Loads on School Children: A Critical Narrative Review

Background: Backpack loads of school students during school days have been suggested to range from 10% to as high as 25% of their body weight and may have a negative impact on their body. The aim of this review was to identify and review studies that have examined impacts of contemporary backpack loads on school children. Methods: A systematic search was conducted of the literature using key search terms. After relevant studies published in recent years were selected using strict inclusion and exclusion criteria, the studies were critically appraised and relevant data were extracted and tabulated prior to conducting a critical narrative synthesis of findings. Results: Twenty-one studies were included, ranging in methodological quality from poor to good (critical appraisal scores 22% to 77%). Students carried on average over 15% of their own body weight, which caused biomechanical and physiological adaptations that could increase musculoskeletal injury risk, fatigue, redness, swelling and discomfort. Conclusion: Considering the limited methodological quality and variations in foci across studies, further research is needed to elucidate: (1) the loads students carry around on a school day in their school backpacks and; (2) the biomechanical, physiological and physical effects of load carriage on students.


Introduction
A review by Mackenzie et al. [1] in 2003 of backpack loads carried by school students during a school day identified that children were carrying as much as 30% to 40% of their body weight. This review, while acknowledging that no critical maximal load had been established (to address back pain), recommended around 10% of the child's bodyweight as a maximum limit. The following year, a review by Brackley and Stevenson [2] stated that the majority of work considered the loads carried by children to be above recommended limits, likewise recommending a maximal load of between 10 to 15% of the child's bodyweight. Since these reviews, more recent research has suggested that these loads are lighter and in some instances may be meeting this recommendation, with loads ranging from 10% [3][4][5][6] to 25% [7][8][9][10][11] of the school child's bodyweight. However, this recent research, in agreement with the earlier reviews, also suggests that these loads have a negative impact (e.g., increased forward lean, pain, skin pressure) on children's bodies [3][4][5][6][7][8][9][10][11].
Items carried by students in their day-to-day school bags have been found to include, but are not limited to, laptops, books, pencil cases, scientific calculators, sports uniforms, school day uniforms and sport-specific training clothing, up to three types of shoes (sport-specific footwear, school leather shoes, and general runners), lunch boxes and full water bottles [12]. To address the requirement to carry backpacks weighted down by all these items, the review by Mackenzie et al. [1] recommended

Search Strategy
To identify and obtain relevant original research for this review, key literature databases were systematically searched using specific keywords relevant to the topic. The selection of keywords was guided by a review of terms used in related relevant published articles. To improve the relevance of search results, filters that reflected study eligibility criteria were applied to each database. The databases were searched for specific key words (see Table 1).

Screening and Selection
Articles identified from the database search were first screened by title and abstract to identify and remove duplicates and articles that did not meet the review's inclusion criteria (see Table 1). Publications reporting studies that used the same data set, analysed in similar fashion, as another study were treated as duplicates and were removed, along with duplicates arising from searching two different databases. All remaining articles were then obtained in full text and subjected to the review's inclusion and exclusion criteria to determine their eligibility for inclusion in the review, identifying a final set of eligible articles for inclusion.

Inclusion and Exclusion Criteria
Inclusion criteria for this review were as follows: (a) human subjects, (b) English language, (c) children 6-18 years, and (d) publication date within the preceding 15-year period. Subject matter experts were consulted to decide upon the 15-year date range employed for this review. Reasons for this time period included the emergence of the use of the traditional (non-touch screen) notebook computers in the classroom and the use of personal laptop computers, which saw a dedicated surge around 2004 [16], with US school districts subsequently encouraged to provide students with personal laptops in 2010 [25].
Exclusion criteria were as follows: (a) studies that addressed backpack carriage in children with disease and illness; (b) studies of backpack usage with non-school children; (c) studies on backpack usage but with no relevance to school children (e.g., backpack-back pain complexity and the need for multifactorial safe weight recommendations [26]); (d) systematic reviews on backpack usage; and (e) articles that were not reports of original research (e.g., policies or guidelines on how to wear a backpack correctly).

Critical Appraisal and Data Extraction
Following study selection, the included studies were critically appraised using the Downs and Black checklist [27]. The checklist has 27 items designed to assess the methodological quality of randomized controlled trials and non-randomized studies and identify the strengths and weaknesses of these studies, and has been used in previous reviews [27]. Most of the items are scored on a dichotomous scale, awarding one point for a 'yes' answer and zero points for a 'no' answer or where the answer cannot be determined from the published report of a study. Item 5 on the checklist, however, is scored on a three-point scale, awarding two points for 'yes', one point for 'partially', or zero points for 'no' or 'unable to determine'. The final question in this checklist, which assesses the statistical power of the study, is normally scored on a scale of 0-5 based on the study's sample size relative to requirements to ensure adequate power. This question was modified for the current review to give one point for a 'yes' answer, indicating the authors of the study reported a power analysis, or zero points for a 'no' answer, indicating the authors did not report a power analysis. This modified approach to the checklist has been used previously to limit subjectivity in scoring [28]. Through this modified approach, the maximum possible raw score became 28, as opposed to the original maximum score of 32, and the raw score for each included study was converted to a percentage score.
All papers were independently scored by two reviewers (Michelle Perrone and Robin Orr), with the level of interrater agreement determined via a Cohen's Kappa analysis. Final scores were determined by consensus between the two raters. Where agreement was not achieved through consensus, a third reviewer (Rodney Pope) determined the final score.
Kennelly's [29] grading system awards a rating based on the Downs and Black raw score given by the raters. However, as Question 27 was modified for this review, the Kennelly grades were modified and presented as percentages to accommodate the modified Downs and Black checklist total raw score (now out of 28 and subsequently converted to a percentage score). On this basis, the grading system applied was as follows: >61% was graded as 'good' quality, 45-61% as 'fair' quality, and <45% as 'poor' quality. This modification has been previously used in critical reviews [30].
Once the critical appraisal of the studies was completed, pertinent data were extracted from the included studies and tabulated. Information extracted from the studies provided a concise and systematic overview of key attributes of, and findings from, included studies. The key table headings were: (a) lead author, title and publication date, (b) demographics of the participants, (c) load conditions, (d) outcomes measures used, (e) main findings and (f) the critical appraisal of studies (CAS), expressed as a percentage. A critical narrative synthesis of findings from the included studies was then conducted.

Results
The results of the search, screening and selection processes are shown in Figure 1 with key outcomes extracted and compiled in Table 2. approach to the checklist has been used previously to limit subjectivity in scoring [28]. Through this modified approach, the maximum possible raw score became 28, as opposed to the original maximum score of 32, and the raw score for each included study was converted to a percentage score.
All papers were independently scored by two reviewers (Michelle Perrone and Robin Orr), with the level of interrater agreement determined via a Cohen's Kappa analysis. Final scores were determined by consensus between the two raters. Where agreement was not achieved through consensus, a third reviewer (Rodney Pope) determined the final score.
Kennelly's [29] grading system awards a rating based on the Downs and Black raw score given by the raters. However, as Question 27 was modified for this review, the Kennelly grades were modified and presented as percentages to accommodate the modified Downs and Black checklist total raw score (now out of 28 and subsequently converted to a percentage score). On this basis, the grading system applied was as follows: >61% was graded as 'good' quality, 45-61% as 'fair' quality, and <45% as 'poor' quality. This modification has been previously used in critical reviews [30].
Once the critical appraisal of the studies was completed, pertinent data were extracted from the included studies and tabulated. Information extracted from the studies provided a concise and systematic overview of key attributes of, and findings from, included studies. The key table headings were: (a) lead author, title and publication date, (b) demographics of the participants, (c) load conditions, (d) outcomes measures used, (e) main findings and (f) the critical appraisal of studies (CAS), expressed as a percentage. A critical narrative synthesis of findings from the included studies was then conducted.

Results
The results of the search, screening and selection processes are shown in Figure 1 with key outcomes extracted and compiled in Table 2. Figure 1. Preferred reporting items for systematic reviews and meta-analyses (PRISMA) diagram [31] detailing results of the search, screening and selection processes. Preferred reporting items for systematic reviews and meta-analyses (PRISMA) diagram [31] detailing results of the search, screening and selection processes. This study recognized schoolbag-related discomfort with the association of individual, physical and psychosocial factors. Carried bag only short distances and duration, thus limiting the exposure to load carriage as primary students leave bag in classroom all day. No physical factors were significantly associated with discomfort, but students did perceive that if a schoolbag was heavy it was also associated with an increased prevalence of back discomfort (p < 0.05). This study's aim was to assess postural parameters in the sagittal plane for an uneven/unbalanced backpack load equal to 10% of a child's body mass. Increase in lumbosacral region angles with carriage of asymmetrical load (p < 0.054). Significant flattening of Thoracic Kyphosis with right shoulder load (p < 0.040). 10% BW load → increase in upper Thoracic spine curvature and moving head forward thus resulting in the flattening of Thoracic Kyphosis (p < 0.040).

70% = Good
Hong and Cheung (2003) Gait and posture responses to backpack load during level walking in children [8].
Y4 and Y5, Primary school males in Hong Kong. Age 9.43 years Height 1.34 cm 11 males total Loads at 0%, 10%, 15% and 20% of student's BW Data was collected in a controlled testing environment (university gym), using a Latin Square design. Four sessions used to collect the data, students walked with set loads for each session (0%, 10%, 15% and 20% of student's BW) for 23 laps of a basketball court (1978 m).
The aim of this study was to explore the biomechanical stresses of continuous load carriage upon children by examining the adaptations of stride and temporal parameters, and trunk posture in the school setting. Gait pattern was not altered by load but a significant increase in trunk inclination (p < 0.05). 20% of students BW load → significant forward lean of trunk especially as walking distance increased (p < 0.05).

56% = Fair
Hong and Li (2005) Influences of load and carrying methods on gait phase and ground reactions in children's stair walking [9]. One-strap athletic bag (across right shoulder) Double strap backpack (on both shoulders) Loads at 0%, 10%, 15% and 20% of student's BW Data was collected in a controlled testing environment (university gym) using a Latin Square design. An in-shoe pressure measurement system (Novel Pedar System, 99 force sensors) was used to record the temporal and kinetic data during stair walking. Students gait was assessed during 400 m flat ground walk with 20% of the student's BW. Students then stair walked 33 steps up and down 3 consecutive times, while carrying 20% BW load.
The purpose of this study was to look at the impacts of load and backpack carrying methods on ground reaction force and gait temporal characteristics during ascent and descent stair walking in children. No significant difference between left and right foot. Athletic bag increased peak force on left foot more than right foot. Load of 15% and above of the students BW showed to induce a significant increase in stance and double support duration with the backpack on (p < 0.05). This study researched the school-bag weight of primary school pupils (aged 7-9 years) and to identify the number of students that carried backpacks in excess of the recommended limit of 10% of their BW. No statistically significant differences in bag weights were identified (p < 0.647). Bag weight varied during the week.Heaviest bag: Y1-5.5 kg (p < 0.001), Y2-7.0 kg (p < 0.001), Y3-6.2 kg (p < 0.001). Carrying backpacks at load of 15% student's BW.
Data collection was done in a controlled environment (university laboratory). Kinematic parameters of the subject's gait at a self-selected speed was analyzed using a six-camera motion analysis system used, with two force plates. Markers located on student's bodies (Helen Hayes, 1991 protocol) [34]. Students were analyzed walking with no backpack vs. walking with backpack on one shoulder at a load of 15% BW.
The aim of this study was to research the biomechanical changes during walking with asymmetric backpack carrying in adolescents, focusing on the effects of asymmetrical backpack carriage on kinematic parameters of the lower body under both loaded and unloaded conditions. When carrying load, and relative to unloaded gait: ankle peak dorsal flexion increased on unloaded side and decreased on loaded side (p < 0.05). Mean knee varum value increased on unloaded side but decreased on loaded side (p < 0.05). In the sagittal plane, knee flexion increased at initial contact on loaded side relative to both the unloaded side and unloaded walking (p < 0.05). Hip joint maximum extension angle decreased on the loaded side compared to unloaded (p < 0.05). Mean hip adduction increased on loaded side and decreased on unloaded side (p < 0.05). Mean anterior pelvic tilt during stance increased on loaded side (p < 0.05). Pelvis was elevated on loaded side and depressed on unloaded side (p < 0.05). Both loaded and un-loaded sides were affected by asymmetrical backpack carriage. Extra load on lumbar vertebral joints and altered frontal knee biomechanics → increase in back pain and pathologies in the knee joint. This study investigated the outcome of backpack carriage in primary school children from the point of view of possible changes occurring in the foot-to-ground relationship. Foot-ground contact was significantly affected by the backpack presence (p < 0.01). Significant effect of the backpack only on forefoot and midfoot regions (p < 0.01). Load up to 10% BW increased plantar pressure in the midfoot and forefoot. Heavy load and exposure time → increase in foot discomfort. Students own backpack mean weight 5.2 kg. Load 15% of student's BW.
Students removed their shoes, then stature, BW, and backpack weight were all recorded. Static plantar pressure distribution measured in upright stance in quiet controlled conditions was acquired (using a pressure platform). Students were asked to standing still for 10 s then asked to walk with and without their school backpack.
The purpose of this study was (1) to evaluate the effects of backpack carriage on planter pressure quantity and issuing, and on spatio-temporal parameters of gait; (2) to examine the association between carried load and plantar pressure parameters. Spatio-temporal gait not affected by load Significantly increased contact areas in the forefoot, midfoot and rear foot (p < 0.001). Significant increase (up to 25%) in plantar pressure during both standing and walking on the fore foot (p < 0.001). Students were divided into two groups: (a) Students who carried backpacks and group; and (b) Students that carried single strap backpacks.
The manner in which the bag was carried was then divided into three categories: 1. Students that carried bags over both shoulders; 2.
Students that carried bags over one shoulder; 3.
Students that carry single strap over one shoulder.
Pain was assessed via a questionnaire (open and closed-ended questions). Students experiencing pain were then divided into those who's bag mass exceed 10% of the student's BW. The students bags were then weighted using calibrated digital bathroom scales.
This study focused on the relationship between school bag carriage and pain in school children.
Examined bag type, individual characteristics, the load carried and the pain experience. Type of school bag, the manner it is carried in and the gender of the student all associated with level of 'Shoulder pain' (p < 0.001).

70% = Good
Students removed their shoes and were asked to stand on a force plate, where their weight was recorded before any measurements were taken. Students then stood on a stadiometer and their height in centimeters was recorded. Image Tool version 3.0 digitizing software (University of Texas Health Service Centre, San Antonio, TX, USA) was used for analyzing photographs.
The aim of this study was to explore the changes in postural angles under various backpack loads in preadolescent children. Craniovertebral Angles changed significantly after carrying 15% BW load (p < 0.05). Head and Neck angles changed after carrying 10% load BW (p < 0.05). Trunk and lower limb angles changed after carrying 5% BW load (p < 0.05). Backpack load changed all body angles and affected overall posture. Students were asked questions through the use of a body chart, pain intensity (using a face pain scale-revised) frequency and consequence of back pain (questions asked: bag type, how the bag was carried, the use of lockers, participation in sport, presence of back pain, pain location). Students weight and bag weight was recorded using stadiometer scales. Students also underwent a face-to-face interview with a physiotherapist.
The purpose of this study was to assess the presence of back pain in school children, as well as its link with school bags. Over 70% had a backpack that exceeded the 10% BW level. 32% complained of back pain, but 74% said it was low intensity pain (p < 0.001). Self-reported pain in school children is independently linked to carrying heavy schoolbags (p < 0.001). Data was collected in a laboratory over four trials with different backpack loads for each student. Students were told to wear specific clothes for the trial (black shorts/tights and no shirt). Disposable surface electrodes were attached to the students right side of their body (upper trapezius and rectus abdominis). Students walked on a treadmill for 20 min with each of the 4 loads.
The aim of this study was to explore the effects of lengthy load carriage on muscle activity and fatigue in children when walking. Increased muscle activity was recorded at all loads (p < 0.05). 15% BW load significantly increased muscle activity in upper traps (p < 0.05). During prolonged walking, the 20% BW load was associated with the most significant muscle activity in upper and lower traps (p < 0.05). Fatigue in upper traps was identified within 10 min under load and in lower traps within 15 min (p < 0.05). No increased muscle activity found in rectus abdominus at any load or any duration of walking (p < 0.05).

63% = Good
Vieira (2015) Impact of backpack type on respiratory muscle strength and lung function in children [38]. Test done in controlled setting at the student's school, using their own school bags. Height and weight was recorded using stadiometer and a scale). Lung function and respiratory muscle strength data was measured in the following settings: (1) Unloaded erect standing position (without the backpack). (2) Bilateral shoulder strap carried over both shoulders (with 15% of the students BW). (3) Bilateral shoulder strap carried over one shoulders (with 15% of the students BW). (4) Padded, adjustable mono shoulder strap.
Inspiratory and expiratory muscle strength was tested with a digital mouth pressure meter, student exhaled slowly then inhaled with force for two seconds (repeated 3 times) The aim of this study was to examine the effects of the backpack type on students' lung function and strength of inspiratory and expiratory muscles in children (aged between 10 and 12). Mono strap restricted and affected lung function, decreasing expiration and muscle strength (p < 0.001). Double strap backpack a preferable option (p < 0.001). Walking with a backpack more than 10% BW increases trunk forward lean and increases breathing rate and decreases trunk Range of Movement (p < 0.001). Student's own backpack Student's shoes were removed and they were weighed as well as their backpacks (using two calibrated digital electronic scales). The content of the backpack was not assessed. Students were then sked 22 questions in an administered questionnaire (chronicity, prevalence, severity and frequency of back pain).
This study identified the associations between school backpack weight (recorded as a percentage of the student's overall BW), how the students wore their backpack, how long they carried the backpack, their socioeconomic status, and the prevalence, severity and chronicity of back pain. Students who walked to and from school and method of wearing the backpack were associated with level of pain. 64% reported back pain. 41% felt pain whilst carrying backpack. Girls reported a higher level of back pain. 77% = Good Goodgold et al. (2002) Backpack use in children [15].
Y5, Y6, Y7 and Y8 primary and middle school students in Boston USA. Age 11-14 years 169 males and 176 females (345 students in total) Student's own backpack.
Questionnaire was used to gather information (student demographic, play and leisure activity levels, bag type and carrying methods). Students were weighed with and without their normal school backpack. Students reported they did not carry their backpacks around with them on a school, they store them in lockers.
The purpose of this study was to describe backpack use by children to assess the severity of the problem. The study examined: (1) Common backpack type used.
(2) Average backpack load.  This study investigated the effects of different backpack carrying methods on school-aged children. No significant difference found during loaded walk with student's base of support, stride length and velocity when compared with the unloaded walk. Double limb support significantly increased with the loaded walk (little difference between one strap carry or two strap carry). Little change in temporo-spatial gait parameters with 15% BW load when compared with no load.

Nature of the Load Carriage
When considering the nature of the loads being carried, all but four studies [9,32,33,40] involved students utilizing their own school bags, whereas the four [9,32,33] divergent studies used dummy/commercial/mono backpacks (Table 2). Four studies observed students carrying backpacks around the school on a daily basis [3,8,9,35], two studies stated students used lockers at school or left bags next to their classroom desk [5,15], one study stated students only carried backpacks over a short distance [7], and five studies did not state the nature of the backpack carriage at school [4,6,18,32,33]. It is important to state that none of the included studies investigated and itemized the contents of the loads carried by students in the real world (i.e., school environment). This lack of itemization of loads is a gap that demands further research, especially given technological advances and the fact that the school students' backpack loads are evolving as a result.

Biomechanical Outcome Measures
A variety of biomechanical impacts of load carriage were investigated ( Table 2). These biomechanical impacts of load carriage included changes in postural parameters and changes to gait parameters.

Posture-Related Biomechanical Measures
The impacts of backpack load on posture focused primarily on changes in postural angles. Drzal-Grabiec et al. [4] examined postural parameters, including habitual posture, with the backpack on the participant's left and right shoulders. The student participants, in this study, exhibited a significant reduction of thoracic kyphosis during load carriage when compared to no load being carried (p < 0.05). In particular, the angle of thoracic kyphosis increased between measurements (p < 0.05) [4]. Similarly, Ramprasad et al. [10], who also examined changes in various postural angles with different backpack weights, found that the craniovertebral (CV) angle changed significantly after 15% BW was reached in backpack loads (p < 0.05). Head on neck (HON) and head and neck (as a single unit) on trunk (HNOT) CV angles were found to significantly change after a load of 10% of BW was reached (p < 0.05) [10]. However, trunk and lower limb angles, which also altered significantly, changed after only a load of 5% of BW was reached in the backpack (p < 0.05) [10]. Furthermore, in a study by Chow et al. [32] changes in spinal curvature and repositioning of the student's centre of gravity (CG) were observed, when backpack loads exceeded 15% of the student's body weight, at different CG locations (anterior and posterior at T7, T12, L3) [32]. This is the same percentage of body weight found to induce biomechanical changes that were reported by Ramprasad et al. [10].
Pau et al. [6] assessed the effects of backpack load carriage on plantar pressure distribution and spatio-temporal gait parameters among children. Measures were obtained during both quiet standing and walking, both with and without the backpack in 208 participants. More than half of the children in this study exhibited a backpack/body mass ratio higher than 15%. Pau et al. [6] found that backpacks introduced significant increases in overall center of gravity (CG) contact area (up to 10%). They also observed that plantar pressure peaked in the forefoot and midfoot regions when loads were carried by children, with a significant shift in the average position of the center of pressure towards the forefoot and midfoot, indicating the body's attempt to restore the initial balance conditions challenged by the load. Pau et al. [6] suggests that heavy loads and significant exposure may increase the risk of foot discomfort and onset of foot structure alterations or pathologies.

Gait-Related Biomechanical Measures
Six studies [3,6,9,11,37,39] investigated the impact of backpack loads on parameters of gait. Three of the studies [9,39], compared both unilateral (single shoulder backpack carry) and bilateral (two-shoulder backpack carry) loads, whereas one study [37] looked at a unilateral load only, comparing a loaded side to an unloaded side. In the study by Hong et al. [9], the researchers examined the effects of carrying methods (backpacks and single-strap athletic bags) and loads on phases of gait and ground reaction forces during stair ascent and descent in children. The threshold load that caused a significant increase in the peak forces observed, regardless of bag type or stair mode, was 15% of body weight [9]. Hong et al. [11] demonstrated that stride length decreased in the right and left legs when the backpack was worn over one shoulder but increased in both legs when the backpack was worn over both shoulders.
Contrary to the findings above, Cottalorda, et al. [3] found that when children carried a backpack (on one or two shoulders) they walked with longer stride, stance and double stance than when walking without a backpack. Connolly et al. [39] found an increase in double limb support when weight was added to the backpack as a counter measure to causing increased instability during gait. They also suggested that when carrying the backpack over one shoulder, the individual may lean in a direction opposite to the force (or load). As such it is not surprising that Pau et al. [6], while not directly observing spatio-temporal gait parameters, observed a significant increase (up to 25%) in plantar pressure, especially in the forefoot.
The consequences of the aforementioned changes in gait parameters from increased load, such as reduced stride length might be explained by the findings of Ozgul et al. [34]. Ozgual et al. [34] found peak ankle dorsiflexion and hip extension, as well as range of pelvic rotation, all decreased in adolescents carrying loads. Furthermore, knee flexion, a loading response for the support phase of gait, increased on the loaded side relative to the unloaded side when loaded walking was compared to unloaded walking. Decreased maximum hip extension during late stance, increased hip adduction, an elevated pelvis and increased anterior pelvic tilt were seen on the loaded side, and if the pelvis was lowered, ankle dorsi flexion increased and the hip was abducted on the unloaded side as a counter effect, when bags were carried on one side [34]. Thus, Ozgul et al. [34] suggest that both unloaded and loaded sides were affected by asymmetrical backpack carriage, putting more load on the lumbar vertebral joints, altering frontal knee biomechanics and contributing to lower back pain. These findings however do require further research given the 'fair' methodological quality (CAS score of 59%) of this research.

Physiological Outcome Measures
Vieira et al. [38] examined the influence of backpack type on lung function and respiratory muscle strength ( Table 2) in both primary and secondary school children carrying a backpack under three controlled conditions: (1) with bilateral shoulder straps used to carry the load evenly on both shoulders, and carrying 15% of the student's body weight; (2) with bilateral shoulder straps used to carry the load over one shoulder, and carrying 15% of the student's body weight; and (3) with padded mono adjustable shoulder straps, carrying the load over one shoulder and across the body, and carrying 15% of the student's body weight. The observed restrictive effect of load carriage and the resulting decrease in expiratory muscle strength were more pronounced for the backpack with a mono shoulder strap, suggesting a double strap backpack is preferable [38].

Measures of Physical Discomfort and Pain
Seven studies (Table 2) investigated physical discomfort imparted by backpacks [5,7,11,13,14,35,36]. The methodical quality of these studies ranged from 'poor' [36] to 77% good [13]. Dockrell et al. [7] analysed musculoskeletal discomfort among primary school students and found that the prevalence of baseline musculoskeletal discomfort was high (63.4%) and that schoolbag-related discomfort was reported more frequently in the shoulders (27.3%) than in the back (15%). The dose-response assessment indicated that both statistically and practically significant increases in discomfort were observed following school bag carriage. Multiple logistic regression models indicated that psychosocial factors and a history of discomfort were predictors of schoolbag-related back discomfort, while gender (being female) and a history of discomfort were predictors of schoolbag-related shoulder discomfort. Puckree et al. [14] also found through their research that more females than males (91.9% vs. 78.5%) experienced school bag carriage pain [14].
Four of the included studies [13,14,35,36] investigated the impact of pain felt by the students from carrying their school backpacks. Two studies [13,14] focused on non-specific mechanical back pain, one study [13] looked at the regions of the body that were most affected by school backpack carriage, one study [35] analysed the influence of backpack weight on back pain and back pathologies, one study [36] analysed backpack injuries in school children and one study [11] analysed the patterns of shoulder and abdominal muscle activation during prolonged walking with loads in children. These studies found that back pain was highly prevalent in the school children, as students reported severity and chronicity of pain was high, with Siambanes et al. [13] finding that 64% of the students reported back pain, 41% felt pain when carrying their backpacks and almost all students reported relief upon taking their backpacks off their backs. Among these children, 13% rated the pain as being 'not bad' and 87% reported the pain to be 'bad/very bad'. Furthermore, 16% reported that they missed days off school, gym classes and participating in sport because of the pain, while 17% of the students reported that they had seen or continued to see a doctor due to their backpack-related pain. The study by Siambanes et al. [13] found 86.9% of the school children participating in their study experienced pain in the region of the back, shoulders and neck, with most of the students relating their pain to lugging heavy schoolbags everyday [13].
In the study of Puckree et al. [14], of the 39 students, 14 reported severe pain and 20 reported moderate pain whilst carrying their backpack on both shoulders. Shoulder pain was reported most commonly, with 15% of students reporting pain in this region of the body, followed by shoulder and back pain at 7%, then neck and shoulders at 5% [14]. Rodriguez-Oviedo et al. [35] analysed the influence of backpack weight on back pain and back pathologies. 61.4% of students carried school backpacks exceeding 10% of their body weight and 18.1% exceeded 15% of their body weight; 25.9% reported having back pain for more than 15 days in the previous year, and the most frequent pathology from these reports was scoliosis (70% reported) followed by lower back pain ad contractures (10% each). Students carrying backpacks weighing in the highest quartile of loads relative to body weight had a 50% higher risk of experiencing back pain for more than 15 days than those in the lowest load quartile, with girls reporting a higher risk of back pain and an increasing risk of experiencing pain with age. It should be noted that in this study, the relative quartile loads were not provided.
Soares et al. [36] analysed backpack injuries in Indian school children, finding pressure marks (redness and swelling) over the neck and shoulders, corresponding to locations of the straps of the backpack, stooping posture while carrying the backpack, pain or stiffness in the neck, upper back and shoulders predominantly while carrying the backpack, and an absence of these symptoms during the school holidays. Soares et al. [36] also found the upper back (40%), neck (27%) and shoulders (20%) were the most prevalent body regions in which pain was reported, followed by the forearm and wrist at 7% and lower back at 6% [36]. Hong et al. [11] found that when assessing the patterns of shoulder and abdominal muscle activation during prolonged walking with loads of up to 15-20% of body weight in children, signs of muscle fatigue were found in the upper trapezius muscles after 10 min.
Five of the 21 included studies (23.8%) used questionnaires [7,[13][14][15]37] asking the students for ratings of pain and body discomfort experienced while carrying their school backpack, through use of pain visual analogue scales (VAS) and ordinal scales. Five of the studies [7,13,33,37] presented results relating to backpack physical discomfort in the wearer. One study by Dockrell et al. [7] asked the student participants to answer questions linked to a body discomfort scale (BDS), and 38% of the students reported baseline discomfort (before they donned the backpack) in their shoulders and 30% reported pain in their back region while carrying the load. Students in the study were also asked questions using the VAS, with results showing a statistically significant difference between scores given before bag carriage (1.6 ± 2.3) and after bag carriage (4.7 ± 2.4) for shoulder discomfort and for back discomfort (before 2.8 ± 2.6; after 5.8 ± 2.4). This study also asked questions related to carrying their school backpack, using a strength and difficulties questionnaire (SDQ). The SDQ questionnaire is a 25-item behavior screening questionnaire, which calculates five dimensions of behavior and emotional state in 4-to 17-year-old children (e.g., emotional symptoms, conduct problems, hyperactivity, peer problems and pro-social behaviors). Parents rate each question on a grading system comprising the grades of 'not true', 'somewhat true' or 'certainly true', with higher scores indicating greater difficulties [7]. From these questions, the researchers found 86% of the students had a normal SDQ score, 8% were borderline, and 6% had an abnormal score [7]. The SDQ score for girls increased as their body weight increased. Another study, using a numerical pain experience scale [14], reported that 87% of students experienced bodily pain in the regions of the shoulders, with 7% feeling pain in the back and 15% feeling pain in the neck [13]. In a study by Spiteri et al. [37], the most common self-reported site of pain in students in grade 5 and 6 (8 and 9 years old) was the neck (25%), and in students in grades 7-9 (10-13 years) was the upper and lower back (44%), with a total of 8% of students in grades 5 and 6 and 17% in grades 7-9 reporting pain in multiple sites in their spines. The prevalence of self-reported back pain was higher among female students, students in non-public schools, and those in secondary schools [33].

Discussion
The aims of this literature review were to identify, critically appraise and synthesise key findings from recent studies investigating impacts of contemporary load carriage on school children, in order to inform future research and assist in the development of risk management for school backpack loads. In total, 21 publications were critically reviewed, achieving a mean ± standard deviation (SD) raw score for methodological quality of 16.86 ± 4.34 (range 5 to 21) out of 28 on the modified Downs and Black checklist [27]. Common weaknesses were identified in the included studies by the Downs and Black checklist [27]. The distributions of principal confounders in each group of subjects were typically not clearly described. External validity was often poor due to the majority of studies using controlled facilities in which to conduct assessments (for example, laboratories, classrooms and treadmills), which were not representative of the environment in which students would typically perform load carriage. For example, they often did not observe walking to and from school, walking around school grounds, walking from classroom to classroom (which may include a number of stairs), activities undertaken during recess and lunch breaks, or the associated requirement to don and doff the backpack. Internal validity also frequently scored poorly, since most of the included studies did not make an attempt to blind the participants or assessors. This was mainly due to the nature of the studies, as it would be difficult to blind participants in the studies given that participants would be aware of either their backpacks or themselves being weighed, and whether or not they were carrying loads at the time of data collection. However, given that many of the papers were rated overall as being of 'good' quality, the findings in general can be considered informative for the present review.
Six naturally emerging themes arose as part of the data extraction. These six themes were: (1) the participant age groups studied and contexts of study have been limited, and this warrants further research; (2) the weight of loads carried were variable but often relatively heavy; (3) the nature of loads carried are unknown (i.e., none of the students backpack contents were itemized in any of the studies); (4) the biomechanical impacts of these loads were statistically significant; (5) the physiological impacts of the loads were also significant but under-researched; (6) the pain and discomfort impacts of these loads were extensive.

Age Groups of Participants
Of the 21 studies included this review, only two (9%) [35,37] focused on secondary school children aged between 13-17 years. It is therefore evident that limited research has been conducted in the secondary school system (ages 13-17 years). This is important to note, given that primary school children typically carry school bags over a shorter distance and, therefore, may have a more limited exposure to load carriage [7]. A study by Mackie et al. suggests that secondary school children typically move from one classroom to another between subjects (according to their timetable) putting on and taking off their schoolbags as required and therefore carry them for longer periods of time [41].

Contexts of Load Carriage That Have Been Studied
Just four of the 21 studies (19%) involved students carrying their backpacks around the school on a daily basis [3,8,9,35]. Only one study (5%) [35] investigated load carriage in secondary school students carrying loads at school on a daily basis. With respect to the contexts of measurements and assessments used in the 21 studies included, in 17 studies (81%) measurements and assessments were conducted in controlled environments (e.g., laboratories, rooms and on treadmills [3,4,6,[8][9][10][11]18,[32][33][34][35][36]38,39] and 15 (71%) reported testing conducted with students standing still [4][5][6][7][8]10,[13][14][15]32,[34][35][36][37][38]. No studies involved measurements or assessments conducted in the school yard or looked at the distance travelled between classrooms or during lunch periods. With research investigating load carriage practices, noting that the contexts in which loads are carried (including the speed, terrain grade and type and distances) [5] can have a greater impact on the carrier than load weight alone, it is imperative that future research includes variation in these contextual factors as part of investigations into school child load carriage. Evaluation of the exact distances travelled whilst carrying their backpacks during a school day is needed. This may include taking into account walking to and from home, moving from classroom to classroom, movement during recess and lunch breaks and the movement after school (e.g., moving to sporting activities/venues/work/home). In addition, the types of terrain should be considered (e.g., concrete, grass or stairs), as should the speed of load carriage (e.g., walking to a classroom, running across an oval following lunch, etc.). These future research requirements were acknowledged by Lasota et al. [5] who suggested that the failure to measure the time a student spends carrying their backpack prevents the determination of the duration of load experienced, which should be considered in future studies.
Eighteen of the included studies involved students utilising their own school backpacks (double/two strapped backpack) [3][4][5][6][7][8]10,11,[13][14][15]18,[34][35][36][37][38][39]. Puckree et al. [14] found that the manner in which a backpack is carried (for example, carried over both shoulders or over one shoulder) significantly affected the number of children who reported pain, and so this should also be considered in future research. In that study, pain also varied with duration of load carriage, another factor that should be considered in future studies on this topic [14]. Hong et al. [9] found that no significant difference between bag types in the impacts of load carriage, with one type assessed being an athletic bag and the other a backpack, but the effects of bag type should be further investigated in future research. Typically, school backpacks lack the rigid frame of an outdoor-style backpack and include only a few pockets in the front, in addition to the main storage compartment [42]. While traditionally very simple in design, school backpacks are often made with padded shoulder straps and backs as well as additional reinforcement to support the intended load [42]. To decrease injury and improve comfort, experts recommend that children use backpacks that match the size of the child [42]. This constrains the number of items that can be carried in the backpack, and, as such may not be feasible without changes to loads needing to be carried. Backpack features from most manufactures can include elements like wide, padded shoulder straps for comfort and to distribute weight across the shoulders, padded back sections for comfort and protection, and multiple compartments for distribution of load [43]. However, these features may also add weight to the backpack itself. One area that was not discussed in the literature was the use of hip belts. A review of load carriage found that the use of hip belts affected energy costs and injury risks in tactical populations [44]. As such there may be value in investigating the use of hip belts in school bag carriage. Likewise, further research into backpack design (e.g., as mentioned above: wide, padded shoulder straps for comfort, padded back sections for comfort and protection and multiple compartments for distribution of load) is needed, particularly as materials from which the backpacks are made evolve.
Of the 21 included studies, only one paper by Pau et al. [18] addressed or mentioned alternative carrying methods (i.e., trolley bags or other kinds of bags) and this was due to these forms of bags being excluded from the study. Therefore, further research is needed to address this gap in the existing literature. In addition, several studies [44][45][46][47] have indicated that high load placement results in significantly higher levels of muscle activity than lower placement of backpacks. The differences in muscle activity are primarily due to the movements and forces arising from the angular and linear acceleration of the trunk whilst carrying load [43]. As such load placement should be considered in future research on load carriage in school children.
A study conducted by Grimmer et al. [45] also discovered that, for postural efficiency, loads should be limited to 10% of body weight and the backpacks should be worn high on the spine, considerations in agreement with load placement in the military [44]. Furthermore, the findings from this study led the authors to conclude that neither age nor gender were a significant factor when comparing postural responses to backpack loads or conditions.

Natures and Weights of Loads Carried
Eleven studies [5,6,8,10,11,14,18,35,37,38] weighing both the students and the backpack, reported bag loads ranging from 5 kg [37] to 7.0 kg [5]. Three studies [9,13,35] reported weighing the backpack only, with loads ranging from 0% (no load) [9] to 20% [9] of body weight. No studies were found that itemized the backpack contents (e.g., books, notebooks, stationary, toys, drinks, lunchboxes, additional uniforms, musical instruments, sporting equipment, computers etc.) nor did any study weigh an empty backpack to determine the weight of the load carriage device (i.e., backpack) itself. Lasota et al. [5] discussed the limiting impact of not investigating backpack contents, making it impossible to determine which specific elements influence the final backpack weight. In a study investigating load carriage with adult police officers, a review by Orr et al. [20] found that the average police officer carries around 7-10 kg of load (10-12% of body weight) and specialist police (like SWAT) can carry loads of around 22 kg (around 25% of body weight) as part of their occupation. While the absolute loads carried by the children in the studies included in the present review may be lighter (i.e., up to 7 kg), children have a lower total body mass than adults and as such it is evident that carried loads will be relatively heavier. As an example, Lasota et al. [5], in a study of Grade 2 children, found that the students (mean age = 8 years) were carrying up to 24% of their body weight (range 1.5-7.0 kg)-a relative load greater than those carried by general police officers and similar to those carried by elite police. This high relative load is also supported by the study conducted by Buhagiar et al. [48], who found that male children in their study (mean age = 9 years) were carrying loads of more than 20% of their body weight in the school environment.
Considering these relative loads, 14 studies in the current review reported testing backpack loads of at least 10% (range 0%/no load to 25%) of a student's body weight [3][4][5][6][7][8][9][10][11]18,[32][33][34]38]. Nearly all studies reported that a load of 10% of the student's body weight was the appropriate maximum weight to be carried by students to limit the effects of load discomfort and injuries in agreement with the earlier recommendations by Brackley et al. [2] of a load of 10 to 15% of the child's body weight. Conversely, Negrini et al. [21] found that around 35% of 11-year-old children in Italy carried backpacks as heavy as 30% of their body weight, stating that the average daily load of 9.3 kg carried by the school children would equate to a daily average of load of 17.6 kg for an 80 kg man. As such, determining and ensuring the optimal backpack load, relative to the student's body weight, is a potential approach that could help mitigate the impacts of load carriage. One potential limitation of an approach of enforcing a maximum relative load, however, akin to the limitation found by militaries attempting to employ a relative loading system or protocol [49], is the requirement for a given number of items to be brought to school by the student. On this basis, if students are required to bring and carry key items (e.g., laptops, required textbooks, etc.) to school, and the weight of these items leads to a load more than the relative load recommendation, meeting a relative load recommendation may not be viable.
Harsha and Berenson [50] suggest that if students were to stop carrying a school bag they could be limiting their exposure to physical activity and even the benefits of daily resistance exercise imposed by the weight of their bag. Shultz et al. [51] go on to state that this is particularly relevant in the current situation of rising levels of obesity in children. Schools have seen a reduction in the duration and frequency of physical education classes and sport in recent times, with the battle of curriculum vs. curriculum activities. Dollman et al. [52] discuss this against a backdrop of rising youth obesity, noting that diminishing curriculum time for PE represents a serious public health issue. Dockrell [7] discusses two conflicting outcomes associated with the activity of carrying a school backpack-on the one hand it provides a form of physical exercise, and on the other it can cause physical pain or damage to a child.

Biomechanical Impacts of School Backpack Loads
Six of the included studies [3,8,9,33,34,39] found that backpack load carriage impacts the gait of school children, resulting in significant changes to gait parameters. Hong et al. [8] found that as loads increased to 15% of students body weight, stride length decreased, with a significant decrease in step length (p = 0.047), cadence (p = 0.057) and walking speed (p = 0.004) [9]. Two studies [8,10] reported findings associated with the student's body posture, including an increase in lumbosacral angles, flattening of the thoracic kyphosis, and deepening of the cervical lordosis. Backpack loads of even 5% of body weight can significantly change trunk and lower limb angles and loads of 15% of student's bodyweight resulted in changes in all the angles pertaining to head, neck, and lower limb, affecting overall posture [4,10]. Two of the studies found that when a backpack was worn it significantly affected foot-ground contact area and increased plantar contact pressure, especially at loads equivalent to 25% of the student's body weight [6,18]. As such, Pal et al. [4] recommends that students should be carefully screened for possible 'high load-high exposure time' with a loaded backpack to reduce the risk of foot discomfort or injuries and to allow correct development of the foot structure and functionality during critical stages in a child's physical development [6]. However, future research is needed to investigate links between load and exposure time, to quantify the level at which risk increases, and to determine whether statistically significant findings do in fact translate to clinical significance. This information can then be used to inform maximal load or maximal time dosages for backpack load carriage. Alternative carriage of the backpack by changing the positioning occasionally between anterior and posterior positions might help relieve the effects of the backpack on the spine, as identified by Chow et al. [32] who found that spinal curvature and repositioning errors were affected by backpack anterior-posterior positioning and CG levels.

Physiological Impacts of School Backpack Loads
The effects of backpack loads on the students' respiratory muscle strength (inspiration and expiration) and lung function was reported in a study by Viera and Garrett [40] involving four conditions: (1) unloaded and standing still in erect stance; (2) 15% body weight load in a bilateral backpack stance; (3) 15% body weight load in a one shoulder backpack stance; and (4) 15% body weight load with a mono strap backpack, in erect stance. They found that when walking with more than 10% body weight load there was an increase in breathing rate and a decrease in trunk range of movement. The mono strap backpack, which is positioned diagonally across the body, was found to significantly lower forced vital capacity, forced expiratory volume in one second and maximal expiratory pressure. This bag design was found to restrict respiratory efforts and decrease expiratory muscle strength, so the double strap backpack is recommended as the preferred backpack option [38].

Physical Discomfort Caused by School Backpack Loads
Eleven of the 21 studies included provided evidence that students who were carrying backpacks recorded bodily pain, redness, swelling, fatigue and/or musculoskeletal discomfort in the upper or lower back, upper or lower trapezius, shoulders, neck or forearms, with most of the students relating their pain, fatigue and discomfort to lugging heavy schoolbags around every day-almost all students reported relief upon taking their backpacks off [5,7,11,13,14,[32][33][34][35]. The prevalence of back pain was reported to be higher among female students, students in non-public schools, and those in secondary schools [33].

Summation of Findings
Students carried on average over 15% of their own body weight and exhibited significant biomechanical and physiological impacts of these loads as well as reporting pain, fatigue, skin redness, swelling and discomfort. Considering the limited methodological quality and variations in foci across studies, further research is needed to better elucidate: (1) the loads students carry around on a school day in their school backpacks; and; (2) the biomechanical, physiological and physical effects of load carriage on students.
From the findings in this review it is evident that there have been limited changes or improvements in student's backpacks (in relation to load and impact) since previous reviews on this topic were conducted in 2003 [1] and 2004 [2]. The earlier review by Mackenzie et al. [1] noted that a student's school book bag was reported to weigh more than 15% to 20% of their body weight, was associated with back pain, and when improperly used resulted in changes to the child's posture and gait. The findings of this review were similar, with reported loads ranging up to 20% of body weight, and associations between load carriage and back pain, changes to gait and posture affirmed. Likewise, where the review by Brackley and Stevenson [2] gave maximum load recommendations for school children of 10-15% of body weight, 14 of the 21 studies in this review reported that a load of 10% of the student's body weight was the appropriate maximum weight to be carried by students to limit the effects of load discomfort, injuries and other adverse impacts [3][4][5][6][7][8][9][10][11]18,[32][33][34]38].
As a final consideration, the review by Mackenzie et al. [1] in 2003, noted that large, heavy backpacks were a relatively recent phenomenon, and discussed inadequate numbers of student lockers, less time between classes to get to lockers, larger textbooks, sports bags, musical instruments, and other objects as areas of concern regarding student loads. While the research highlighted in this review bears many similarities to that considered in earlier reviews, the contextualization of load carriage in regards to lockers, additional items being carried and travel between classrooms, and in general, was limited. As such, the load weights carried by students, backpack pain and other concerns have shown little to no improvement over the past 15 years and the recommended load of around 10% body weight is still maintained, although typically still not met. The contextualization of the carriage of these loads is also very limited. Ultimately, the concerns and unease regarding load carriage in school children remain.

Limitations of This Review
The literature search employed in this review was limited to two databases due to the relevance of those databases and the high numbers of papers that were identified from them. Whilst stricter search terms may have limited the overall number of initial studies identified, this approach may have meant that relevant papers were not identified in the initial search. On this basis, while limited to two databases, the search terms used were deliberately kept broad to minimize the risk of not identifying key articles.

Conclusions
Based on the methodological quality of studies included in this review and limitations in the contexts in which these studies have been conducted, it is apparent that further research into the impacts of backpack loads in real-world school settings is required using a wider range of school year levels and student ages. In particular, further research of this nature is required to assess (1) the contemporary loads students carry around on a school day in their school backpacks; (2) the impacts those loads may have on the student's body; and (3) the biomechanical, physiological and physical effects that occur to the students as a result of carrying these loads. Despite the shortcomings of recent, available research in this area, it is nevertheless apparent that the wearing of school backpacks does have significant biomechanical, physiological and discomfort impacts on the wearer, especially with loads above 10% of the student's body weight. Such impacts may include changes to posture (e.g., changes to spinal posture, lumbo-sacral angles, and thoracic kyphosis), gait (increases in plantar pressure during foot-ground contact and increased double support), increases in physical discomfort and muscle activity, and increases in breathing rate. Further research is required to identify and select optimal backpacks for school children, with careful consideration of the loads to be carried on a daily basis, duration of backpack load carriage and how students best carry the load to minimize the negative outcomes associated with wearing school-related backpacks.
Author Contributions: M.P. conceived the critical review topic, performed the systematic search of the literature, critical appraisal analysis, and data extraction, analyzed the data, synthesized the findings, and drafted the manuscript. R.O. confirmed the literature search, conducted a critical appraisal analysis and the Kappa analysis, moderated final critical appraisal scores, and reviewed the manuscript. R.P., W.H., and N.M. aided in the development of the search terms and strategies, aided in elements of drafting the manuscript, and reviewed and edited the final manuscript.
Funding: This research was supported by an Australian Government Research Training Program Scholarship and APC was funded by Bond University.

Conflicts of Interest:
The authors declare no conflict of interest.